Dosu LogoDosu Logo
Ask
Join our Discord
Organization avatar
HelmorPublic
Caspian's Workspace
DocumentsHelmor
scaling-support-and-documentation-at-Dify-one-of-the-fastest-growing-oss-projects
scaling-support-and-documentation-at-Dify-one-of-the-fastest-growing-oss-projects
Type
External
Status
Published
Created
Jul 14, 2026
Updated
Jul 14, 2026
Source
frontend/marketing/data/case-studies/scaling-support-and-documentation-at-Dify-one-of-the-fastest-growing-oss-projects.mdx

Dify's Origins#

Two years ago, we open-sourced Dify, a production-ready platform for agentic workflow development that combines Backend-as-a-Service with LLMOps to simplify generative AI development. Thanks to our incredible open-source community, Dify has since grown into a powerful solution for real-world AI applications. We’ve now reached 100,000 GitHub stars, placing Dify among the top 100 open-source projects worldwide, trusted and used by global leading enterprises.

Dify: Build Production-Ready Agentic AI Solutions

One downside to all this popularity is the amount of Github issues we deal with. On a monthly basis, we see around 500 issues that we had to manually respond to with a team of under 10 people, and as any open source maintainer of even a moderately sized project knows, managing issues and pull requests can be very time consuming.

Enter Dosu#

Issue interactions per day pre & post Dosu

We were looking for a tool that could help triage issues automatically and we stumbled upon Dosu through Langchain, which was already using it to triage issues, so we decided to give it a try.

Since we installed Dosu on our repo and told it to respond to open issues when it can, it has made the process remarkably efficient. We're thrilled to see contributors and users resolve issues simply by mentioning @... in the thread—many even call it “an awesome bot.” What truly impressed us is Dosu’s deep understanding of code. It references multiple data sources to support its responses and always communicates in a polite and helpful manner.

Being deeply involved in RAG ourselves, we know just how hard it is to get it right — and we’re genuinely impressed by the work Dosu has done here.

What's next for Dify: RAG Pipeline 2.0#

We're excited to introduce RAG Pipeline 2.0 - a modularized and more flexible way to transform raw data into intelligent, context-aware responses using LLMs or AI Agents. With a variety of useful plugins on the marketplace, users will be able to orchestrate their data processing pipelines - making it easier to build powerful, domain-specific RAG solutions that truly fit their needs.

This update will empower developers and enterprises to build tailored RAG Pipelines with ease - unlocking innovation at every step.

Conclusion#

If you’re building agentic AI and looking for a platform to simplify development, you should check out Dify. And if you’d like some help managing your product’s issue triage, definitely check out Dosu.

@Dosu
Documents
01 - Get Started
01 - Introduction
02 - Install Helmor
03 - Add a Repository
04 - Create a Workspace
05 - Send Your First Agent Task
02 - Core Concepts
01 - Repository
02 - Workspace
03 - Session
04 - Agent
05 - Changes and Review
06 - GitHub Integration
03 - User Workflows
01 - Start a New Task
02 - Review Agent Changes
03 - Continue an Existing Session
04 - Commit and Ship Work
05 - Use Multiple Agents Safely
04 - Developer Setup
01 - Local Development Setup
02 - Architecture Overview
03 - Frontend Overview
04 - Backend Overview
05 - Sidecar Overview
06 - Testing Guide
07 - Debugging Guide
05 - Troubleshooting
01 - Agent Not Responding
02 - GitHub Authentication Issues
03 - Workspace Errors
04 - Sidecar and CLI Issues
05 - Logs and Diagnostics
7-data-driven-checks-to-ensure-your-documentation-roi-is-high
AGENTS
AGENTS
AGENTS
CHANGELOG
CLAUDE
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
README
SKILL
SKILL
SKILL
STRIPE_README
[tag]
_template
advanced-full-text-search
advanced-jsonb-indexing
ai-contributions-open-source
ai-is-better-at-maintaing-docs-than-humans
april-2025-newsletter-dosu-does-docs
august-2025-dosu-drop
automating-github-issue-triage
better-agent-testing-with-distributed-dependency-injection
blog_template
building-the-future-of-knowledge-management-june-dosu-drop
celery-preserializers-a-low-friction-path-to-pydantic-support
channel-creation-github-repository
channel-creation-slack-channel
ci-jobs
cncf-dosu-case-study
combating-open-source-maintainer-burnout-with-automation
concepts-channel
concepts-data-sources
concepts-interactions
configuration
confluence_edited_markdown
confluence_original_markdown
conn-idle-timeout
conn-limits
conn-pooling
conn-prepared-statements
data-batch-inserts
data-n-plus-one
data-pagination
data-upsert
december-2025-dosu-drop
default
devin
docling-open-source-growth-with-dosu
dosus-new-fact-based-reasoning-agent-is-generally-available
features-auto-labeling
features-generate-docs
features-issue-triage-qa
features-maintain-docs
february-2025-newsletter-introducing-spaces
from-maintenance-to-innovation-freeing-developers-with-ai-automation
george-xu
github-configuration
github-features-dosu-in-first-awesome-continuous-ai-list
github-installation
how-dosu-used-langsmith-to-achieve-a-30-accuracy-improvement-with-no-prompt-engineering
how-lancedb-supercharged-our-knowledge-graph
how-preset-keeps-apache-superset-healthy
how-triliumnext-revitalized-an-abandoned-open-source-project-with-dosus-help
index
index
internal-vs-external-documentation-audience-first
iterating-towards-llm-reliability-with-evaluation-driven-development
january-2026-dosu-drop
july-dosu-drop-public-spaces-enter-the-chat
knowledge-management-in-the-ai-era-why-documentation-is-more-critical-than-ever
li-zheng
lock-advisory
lock-deadlock-prevention
lock-short-transactions
lock-skip-locked
march-2025-newsletter-dosu-knows
marcos
markdown
mastering-auto-labeling-taming-the-backlog-with-intelligent-labels
may-2025-dosu-drop-new-agent-self-serve-pricing-fresh-look
member-invitation
michael-ludden
michael-mangus
migrate-celery-to-dbos-dosu
monitor-explain-analyze
monitor-pg-stat-statements
monitor-vacuum-analyze
notion_edited_markdown
notion_original_markdown
november-2025-dosu-drop
october-2025-dosu-drop
onboarding-introduction
onboarding-key-concepts
onboarding-quickstart
open-source-labeling-best-practices
page
page
pages
photos
pull_request_template
query-composite-indexes
query-covering-indexes
query-index-types
query-missing-indexes
query-partial-indexes
reliable-data-orchestration-for-ai-applications
response-previews
scaling-support-and-documentation-at-Dify-one-of-the-fastest-growing-oss-projects
schema-data-types
schema-foreign-key-indexes
schema-lowercase-identifiers
schema-partitioning
schema-primary-keys
security-privileges
security-rls-basics
security-rls-performance
september-2025-dosu-drop
settings
slack-configuration
slack-installation
sparrowhawk
taylor-dolezal
the-code-understanding-paradox-when-ai-makes-writing-code-fast-but-understanding-it-slow
top-5-zombie-problems-resurrected-by-ai-tools
using-ai-to-generate-and-maintain-documentation
what-lodashs-issue-backlog-reveals-about-oss-burnout
what-the-mythical-man-month-got-right-about-documentation
word2vec-to-transformers-ai-evolution